There is a Latin phrase, "post hoc ergo propter hoc," which translated means, "after this, therefore because of this." This phrase represents a common logical fallacy that if two events happen to occur in conjunction with each other, one must have caused the other. This is not necessarily the case, but this false assumption is responsible for much confusion in science and the social studies. Data can be easily misinterpreted or even manipulated by believing that correlation is the same as causation.
When two variable seem to coincide with each other, we say they are correlated. For example, there is a high degree of correlation between the amount of education you have and the amount of money you can expect to earn in your lifetime. Thus, knowing how educated someone is will allow you to better predict their income than you could by knowing only the average income of the group you are studying, whether that be Latinos, people in their 30s, or all Americans.
The relationship between two correlated variable is usually pretty clear. For example, if you find that men are more likely to support capital punishment than women, you would know that gender influences political ideology. It could not be the other way around. You gender would not vary depending on your political ideology!
Sometimes, however, the relationship is less clear. Sometimes, the correlation could go either way. For example, in our education illustration we said that income depends upon education. That may not necessarily be the case. It could be that wealthier people are more likely to attend college than poorer people. In fact, that is most certainly the case. Education may be more an indication of wealth, not one of its causes.
While we generally assume that variables that correlate have a direct relationship with each other, this is not necessarily the case either. Correlation does not equal causation! For example, there is a high degree of correlation between ice cream sales and murder rates. Are people who purchase ice cream likely to get on some kind of sugar rush and go on a murdering rampage? Of course not! There is another factor involved that we have not yet identified.
Ice cream sales and murder rates correlate because they are both influenced by temperature. Much more ice cream is sold in summer months than in winter months, for obvious reasons. It is less clear why murder rates go up in the summer, but there are several logical factors such as school being out of session, people becoming irritable in the hot weather, and the extra daylight encouraging people to stay out later.
It does not take much thought to realize that ice cream does not cause murder or vice versa, but it is not always so easy to tell if a pair of correlated variables are actually directly related. Many studies you see reported in the news media may begin with invalid assumptions about correlation and causation. As the political season heats up, remain skeptical about statistics you hear reported. If you think critically about a particular study or claim, you may realize that the reported correlation is not actually the result of causation, but of some other factor.
